Synthetic Nucleosides. LIV.lr2 Facile Displacement Reactions in the D

LIV.lr2 Facile Displacement Reactions in the D-Mannitol. Series. I. B. R. BAKER AND A. H. HAINES. Department of Medicinal Chemistry, School of Pharmac...
1 downloads 0 Views 566KB Size
BAKER .4ND HAIIWS

438

Synthetic Nucleosides.

VOL. 28

LIV.lr2 Facile Displacement Reactions in the D-Mannitol Series. I B. R. BAKERAND A. H. HAINES

Department of Medicinal Chemistry, School of Pharmacy, University of Bufalo, Buffalo 14, New York Received SeptembeT 24, 1962 The mesyloxy group of 4-0-benzoyl-1,2 : 5,6-di-0-isopropylidene-3-O-mesyl-~-mann~tol (11) could be displaced by nucleophilic groups, in contrast to most acetonated furanosides and pyranosides containing a ring sulfonate. The mesyloxy group of I1 could be displaced anchimerically by the adjacent benzoyloxy group, which on debenzoylation afforded the C-3 inversion product ( V I ) with the D-talitol configuration. The bimolecular reaction of I1 with acetate ion was shown to be weakly competitive with the anchimeric reaction; the more nucleophilic azide ion was strongly competitive with the anchimeric reaction and allowed introduction of the azido group into the hexitol chain. The anchimeric assistance with an adjacent 4-methoxyl group was negligible, as anticipated, and evidence for S"2 displacement of the mesyloxy group of IX by acetate was observed.

Displacement reactions of a secondary sulfonate on a carbinol to the ketone is the ability of the carbinol to carbohydrate derivative, useful for synt'hesis of rare break a t the C-H bond,ll it follows that the difficulty of displacement of the corresponding carbinyl sulfonate sugars, has been limited primarily to sulfonates on a side chain carbon or to cases where neighboring group (11) with Walden inversion would be roughly parallel. participation can occur. For example, reaction of That such is the case has now been demonstrated in methyl 6-deoxy-2,3-O-isopropylidene-5-O-tosyl-~,~-althe D-mannitol series. lofuranoside with sodium benzoate in boiling dimethylTreatment of 3-0-benzoyl-1,2 :5,6-di-O-isopropyliformamide readily forms methyl 5-0-benzoyl-6-deoxydene-D-mannitol (I) with methanesulfonyl chloride 2,3-O-isopropylidene-cr,~-talofuranoside with Walden inversion3: in contrast, 1,2 : 5,6-di-O-isopropylidene3-O-tosyl-~-glucofuranose, a compound with a ring sulfonate, fails to react under these condition^.^ Only in relatively rare instances of favorable conformation and reactivity can ring sulfonates be displaced by the sodium benzoate-dimethylformamide reagent, such as conversion of methyl 2,3-di-0-benzoyl-4,6-di-0-tosylapgalact'opyranoside to methyl 2,3,4,6-tetra-O-benxo yl- a ,~-ghcopyranoside. although neighboring participation by an acetamido group is a feasible reaction for inversion of carbohydrate configuration,'j trhis type of inversion by the less react,ive7 neighboring acyloxy group has been uniformly unsuccessfu14~s; these results are attribu-t'edt'o the low reactivity of the system which results only in slow cleavage of the acyloxy group, rather than the desired C--0 cleavage of the sulfonate ester. The relative ease of oxidation of the open chain hydroxyl of 3-O-benzoyl-l,2 :5,6-di-O-isopropylideneD-mannitol (I) with chromium trioxide in pyridine to the corresponding ketone, 4-0-benzoyl-1,2 :5,fi-di-Oisopropylidene-~-arabo-3-hexulose, in 507" yieldg contrasts drastically with the poor yields (0.3-5s) obtained by oxidation of a carbohydrate ring hydroxyl.1° Since one of the import'aiit factors in oxidation of a (1) A list of collected references on synthetic nucleosides by B. R. Baker and co-workers will be sent on request. (2) This work was generously supported by the National Cancer lnstit u t e , grant no. CY-5845. ( 3 ) E. J . Reist, L. Goodman, and B. R. Baker, J . Am. Chem. Soc., 80, 5775 (1958). (4) Unpublished work b y B. R. Baker a n d co-workers. ( 5 ) E. J. Reist, R. R. Spencer, and B. R. Baker, J . Org. Chem., 24, 1618 (1969). (6) (a) B. R. Baker and R . E. Schaub, ibid., 19, 646 (1954); (b) B. R . Baker, R. E. Schaub. and J. H. Williams, J . Am. Chem. Soc., 77, 7 (1955); (c) K. W ,Jeanloz, ibid., 79, 2591 (1957). (7) S. Winstein and R. Boschan, ibid., 72, 4669 (1950). (8) S. Peat a n d L. F. Wiggins, J . Chem. SM., 1088 (1938): R. W. Jeanloz Jeanloz. J . Am. Chem. Soc., BO, 5692 (1958). (9) J. M. Sugihara and G. U. Yuen, ibid., 7 9 , 5780 (1957). (10) 0. Theander, et al., Acta Chem. Scand., 16, 437 (1961); 12, 1507 (1958); 11, 1557 (1967); E. Brimacornbe, J. 9. Brimacornbe, a n d B. Lindberg, ibid., 14, 2236 (1960); B. R. Baker and E. J. Reist, unpublished d a t a on chromium trioxide-pyridine oxidation of 1,2:5,6-di-O-isopropylideneD-g~ucofuranose a n d 1,2-0-iso~ro~ylidene-5-carbornethoxy-D-xy~ofuranose.

t-

VI

V

in pyridine gave a 96% yield of sirupy 4-0-mesyl derivative (11), suitable for further transformations; this LTltimately crystallized and had m.p. 73-75'. Treatment of I1 with sodium acetate in boiling dimethyl(11) F. H. Westheimer and N. Nicolaides, J . Am. Chem. S o c . , 71, 25 (1 949) *

FEBRUARY, 19G3

SYNTHETIC NUCLEOSIDES.LIV

formamide containing 0.5% of water for twenty-two hours gave what appeared to be a mixture of monohenzoates (17). One of these two benzoates could be isolated crystalline, m.p. 124-12-5'. Debenzoylation of the crystalline benzoate with methanolic sodium methoxide gave 1,2:5,(i-di-O-isopropylidene-~-talitol (VI), m.p. 62-64', identical with an authentic l2 Debeiizoylation of the mixed benzoates (V) gave crystalline VI in 44% over-all yield from I tia 11. This preparation of VI is stereospecific and is probably a superior preparative method to reduction of the ketone to two diastereoisomers which can be separated by column technics to give about 25% yield of VI from I.9 Since it is possible for V to arise either by direct S s 2 displacement of the mesylate group of 11 by acetate via I11 or by participation of the neighboring benzoyloxy group via the ortho ester ion, IV, the reaction was investigated in greater detail to shed light on the possible mechanism. Chromatographic separation of the crude reaction product gave, in addition to the mixed hydroxy benzoates (V). about 10% yield of an oil having no hydroxyl absorption in the infrared, but showing characteristic acetate bands with carbonyl a t 1725 ern.-' and greatly increased C--0-C a t 1220 cm.-*; the benzoate group showed C=O a t 1720 cm.-' and C-0-C at 1265 em.-'. Deacylation of this acetate with methanolic sodium methoxide afforded the crystalline talitol derivative, VI, showing that the acetate must have structure 111. Since 0acetates can be hydrolyzed under the conditions of this reaction,6b111 could also serve as a source of the monobenzoates, V. Thus the Sr2 mechanism can and does operate. I t was also of interest to see if fluoride ion was a sufficiently strong nucleophile to displace the mesylate of 11, since the resultant fluorohexitol could feasibly be converted into a 2'-fluoro-2'-deoxynucleoside of potential use in chemotherapy. Kot too surprising was the fact that fluorine was not introduced into the hexitol molecule when the mesylate (11) reacted with sodium fluoride in anhydrous dimethylformamide a t the boiling point; in fact, the same crystalline benzoate (V) was obtained as in the sodium acetate experiment, which, in tiirn was debenzoylated to crystalline VI. The formation of V in the sodium fluoride exppriment can only be interpreted to proceed cia the ortho ester ion (IV). JVithout sodium fluoride, the mesylate group of I1 was still ejected in boiling dimethylformamide to give an oil with an infrared spectrum compatible with structure V; however, the crystalline talitol (VI) could not be isolated on debenzoylation indicating that some additional structure changes might have taken place. Thus, the fact that 111could be isolated from the sodium acetate experiment and the fact that the sodium fluoride gave V oia an ortho ester ion (IV) show that both the S N and ~ the neighboring group participation reaction can operate in a molecule such as 11. Therefore, it should be possible to influence the ratio of the Sx2 route to the anchimeric route by using either more powerful nucleophiles or more or less powerful anchimeric groups. In order to investigate the effects of two less effective anchimeric groups, the 0-p-nitrobenzoyloxy and the (12) We wish t o thank Dr. Sugihara for sending us a n authentic sample of this compound.

430

methoxyl groups were investigated. Methylation of I with methyl iodide and silver oxide in dimethylformamidel3 gave VI1 as an oil, which could be debenzoylated with methanolic sodium niethoxide to crystalline VIII14 in 46% over-all yield from I. Treatment of VI11 with mesyl chloride in pyridine afforded 75% of IX, m.p. 72-73'. ( C & )0 z